In the spring of 1986, a diver named Kihachiro Aratake dropped into the sea off the south coast of Yonaguni, the last speck of Japan before Taiwan, looking for a new place to show tourists the hammerhead sharks. About twenty-five metres down the seabed stopped being seabed. It rose in front of him in flat steps, wide as rooms, with edges that ran dead straight and turned in clean corners. He hung in the current and looked at it. Then he swam back up and, by his own account, told almost no one for a while, because he was not sure what he had found or whether saying so would make him sound mad.
Forty years later the question he swam away from has still not been answered. The thing on the seabed off Yonaguni is real, and anyone with a wetsuit and a boat can go and touch it. What no one can prove is the only thing people actually want to know: did human hands make it, or did the sea do all of it alone?
That gap — between a solid object and an unsolved cause — is where the whole story lives. The monument, as the divers named it, has been called Japan’s Atlantis, a ten-thousand-year-old pyramid, the oldest building on Earth, and a perfectly ordinary lump of sandstone that only looks built because our eyes are hungry for corners. All of those claims are about the same pile of rock.
Twenty-five metres down, a staircase that should not be there
Yonaguni sits at the far western end of the Ryukyu chain, closer to Taipei than to Okinawa’s main island. The water there is clear and fast. A strong current, the Kuroshio, sweeps past the coast, which is exactly why the hammerheads gather and exactly why diving the site is not for beginners. The formation lies just off a headland called Arakawabana, its top around five metres below the surface and its base down near twenty-five.
What you see when you drop onto it is a single mass of rock, roughly the length of a football pitch, terraced like a hillside that someone has cut into giant shallow stairs. There are broad flat platforms. There are near-vertical walls that meet those platforms at something very close to a right angle. There is a channel that looks like it was carved to carry water, a pair of tall stones standing side by side, and a curved feature the divers took to calling the loop road. Set all of that in green half-light with fish moving through it, and the mind does the obvious thing. It reads architecture.
That first reading is powerful, and it is worth being honest about why. Human beings are pattern-finders by trade. We are very good at spotting the straight line, the repeated shape, the deliberate corner, because for most of our history those things meant another person had been here. Show someone a right angle underwater and their first thought is rarely “interesting rock.” It is “who built this.” The monument works on exactly that reflex, and the argument over it is really an argument about whether the reflex can be trusted.
The scientist who wanted a city
The man who turned a diving curiosity into a scientific claim was Masaaki Kimura, a marine geologist at the University of the Ryukyus. Kimura dived the site again and again, over years, and came back convinced. To him the platforms were not accidents. They were floors. He mapped the whole mass as a ruin and gave its parts the names of a lost town: terraces and a stadium, a castle, roads, a triumphal-looking gateway, even a stretch he read as a drainage system. He pointed to what he took for tool marks, and to a few loose stones he thought had been shaped and moved.
Kimura went further than the stone. Nearby he identified a rock he described as a carved turtle, and a large feature he read as a human face or head looking out from the reef. He argued the site had been a real settlement, drowned when the sea rose, and at various points he put an age on it — sometimes a few thousand years, sometimes stretching the estimate back much deeper, towards the last time this ground stood in open air. His picture was not a temple built underwater. It was a town that had once been dry, swallowed whole.
The monument is not small and it is not a trick of one bad photograph. The terraced block runs for something over a hundred metres and stands around twenty-five metres from base to top — taller than an eight-storey building. The single largest section is one continuous piece of rock, not a wall of stacked blocks. That last detail cuts both ways in the argument: a monument carved from living bedrock is a staggering feat, but a shape carved by nothing but water needs no builders at all.
Kimura’s work gave the site something the diving stories could not: a named expert, a university, and a body of measurements. It also gave it a champion outside science. The writer Graham Hancock, who has spent a career arguing for a forgotten ice-age civilisation, took Yonaguni up as evidence and put it in front of a global audience. From there the monument stopped being a local curiosity and became a fixture of the lost-world shelf, filed next to Atlantis and the rest.
The geologist who saw only rock
Not everyone who dived it came up converted. The most careful sceptic is Robert Schoch, a geologist at Boston University, who went to Yonaguni expecting, if anything, to be persuaded. He was not. What he found underwater, he wrote, was a textbook example of how a certain kind of stone falls apart.
The rock at Yonaguni is sandstone and mudstone, laid down in flat sheets around twenty million years ago and later tilted. Stone like that has two habits that matter here. It splits cleanly along its bedding planes, the horizontal layers, which gives you flat platforms stacked one above another. And it is cut through by joints — vertical fractures that tend to run in straight, parallel sets. Put the two together and the rock breaks into blocks with flat tops and sheer sides that meet at right angles, all on its own, with no chisel involved. Add a fast current dragging loosened slabs away, and the terraces clean themselves up into something that looks made.
Schoch made one more point that is hard to argue with. The same kind of stepped, angular formation shows up on the coast of Yonaguni above the waterline, where no one claims a lost city, and along other shores in the region. If nature carves these shapes in the open air where we can watch, there is no special reason to insist that the one underwater needed people.
The myth: the Yonaguni Monument is a deliberately built structure — a pyramid or a city — carved by an advanced civilisation and later lost beneath the waves, proof of a sophisticated culture written out of the record.
The reality: the shapes people read as walls, steps and roads are the ordinary behaviour of flat-bedded sandstone that splits along its layers and its joints, and near-identical formations sit above water nearby where nobody claims a builder. There are no confirmed inscriptions, no tools, no worked blocks that have survived independent testing, and Japan’s own heritage authorities do not recognise the site as artificial. The monument is real. The city around it has never been found.
Why the sea was once dry here
Here is the part that gives the wild version its grip, because this piece is not disputed at all. Yonaguni really was above water once. During the last ice age, so much of the world’s water was locked up in ice sheets that the sea stood roughly a hundred and twenty metres lower than it does today. Coastlines everywhere reached out much further. The seabed the monument sits on was open ground, walked on by anyone who lived along that shore.
As the ice melted, the water came back and drowned all of it. By something like ten thousand years ago the rising sea had closed over the ledge at Arakawabana and the monument passed into the dark, where it has been ever since. That much is plain geology. It means the site was standing in daylight at a time when, somewhere across the world, the first people were only just beginning to raise stone circles and temple mounds.
So the timing does two jobs at once. It hands the believers a real, provable window: yes, a person could once have stood on that rock. And it sets a brutal condition on the dream. If human hands shaped the monument, they did the work before the sea rose — which would make it far older than any city we know, older than the pyramids by thousands of years, older even than the great stone rings of Anatolia. A drowned town at Yonaguni would not just add a footnote to history. It would break the story of when we learned to build.
The trouble with a straight line
Everything, in the end, comes down to how much you trust a right angle. The believers point at the corners and the parallel steps and say order like this does not happen by accident. The geologists point at the same corners and say, in this rock, order like this happens all the time.
The sceptics have the stronger hand, and it is worth being clear why. Nobody has pulled a carved block free of the site and shown it was quarried and set. The famous inscriptions dissolve, on close inspection, into natural scratches and the marks left by grazing animals. The “face” and the “turtle” are the kind of shapes we see in clouds. And the single most awkward fact for the lost-city case is that the monument is nearly all one solid piece of bedrock. A carved monument is possible. But a town needs walls made of separate stones stacked by hand, and those are exactly what the site does not have.
None of which quite kills the wonder, and this is why the argument will not die. The believers are not wrong that people were once on this ground. They are wrong, most likely, about what the ground itself is. But “most likely” is not “certainly,” and the sea does not give up its evidence easily. Every diver who goes down comes back with the same first impression Aratake had, and impressions are stubborn things.
What would actually settle it is not another photograph but a spade. If the monument is a ruin, there should be a debris field around its base — broken blocks, worked stone, the rubbish any settlement leaves behind — buried in the sediment nearby. A proper underwater excavation, sifting that sediment for tools, hearths or shaped stone that could be dated, would either turn up a civilisation or turn up nothing. That dig has never been done at the scale the question deserves, which is part of why, forty years on, both sides can still stand on the same rock and see two different worlds.
On paper, just a rock
Officially, Japan has made up its mind, quietly and without drama. The Agency for Cultural Affairs and the Okinawa authorities have never listed the monument as an artificial site or a protected cultural property. In the eyes of the state it is a natural formation, a striking one, and a fine place to dive. No museum holds a Yonaguni artefact, because no accepted artefact exists.
That has not slowed the boats. The monument has become the island’s best-known draw, dived in winter when the hammerheads come and the water is at its clearest. Visitors glide over the terraces and the steps and make up their own minds, which is perhaps the most honest way to meet the place. The science says rock. The eye says stairs. Standing between the two is most of the appeal.
The deeper lesson of Yonaguni is not about a lost race. It is about us — about how badly we want the past to be grander and stranger than the plain record allows, and how readily a hungry eye turns a fracture into a doorway. The sea off Arakawabana kept a genuine marvel: a ledge that people once stood on, ten thousand years before it went under. That should be enough. The strange thing is how often it is not.
Go down there yourself and the argument stops being abstract. The current pushes at you, the light comes and goes, and the great flat steps march away into the blue exactly as if someone laid them. You will want to believe a hand made them. That wanting is the real find at Yonaguni — older than any city, and still not solved.
Sources & further reading
- Robert M. Schoch, Voices of the Rocks and his published field notes on Yonaguni — the geologist’s case that the formation is natural, based on bedding planes, jointing and comparison with formations above the waterline.
- Masaaki Kimura, University of the Ryukyus — papers and conference presentations arguing the monument is a submerged built structure, with his mapping of the terraces, “loop road” and other named features.
- Patrick D. Nunn, Vanished Islands and Hidden Continents of the Pacific — a geologist’s survey of drowned coasts and the science of post-glacial sea-level rise in the region.
- Graham Hancock, Underworld: The Mysteries of a Lost Civilisation — the fullest popular case for Yonaguni as evidence of a forgotten ice-age culture, and the book that carried the site worldwide.
- Japan Agency for Cultural Affairs / Okinawa Prefecture — the official position that the site is not recognised as an artificial cultural property.
- Boston University research summaries on the Yonaguni surveys — overviews of the diving investigations and the geological counter-argument.
Image credits (via Wikimedia Commons): all photographs of the Yonaguni Monument — main terrace, terraces, terrace steps, twin megaliths, loop road, and water channel — by Melkov (CC0).
